Field of Invention
The present invention provides a construction method of a software accelerated testing technology based on a noise resonance, which relates to realization of the software accelerated testing technology based on the noise resonance and belongs to fields of software credibility and software testing.
Description of Related Arts
With the rapid development of the computer technology, the function of the software becomes more and more powerful, meanwhile the scale and complexity of the software increase sharply. In order to decrease the loss caused by the software fault, software testing is made before the official launch of the software. Therefore, it is very important to find the software fault and remove the fault, so as to ensure the quality and reliability of the software. At the same time, because of the sharply increased software scale and the high testing cost due to the time-consuming laborsome conventional software testing method, the requirement of realizing the testing automation becomes increasingly urgent. One of the automatic testing methods, which is relatively mature and popular at present, is the automatic test case generation method based on the evolving method.
The general thinking of the current widely applied evolving method is to adopt the test case set as a population and adopt the coverage rate of the test case as the coverage rate function, so as to realize optimization of the test cases through the respective evolving mechanism. Meanwhile, many other optimization methods are provided, for solving the problem that the evolving method is unavoidable and easy to fall into the local optimum and for continuously increasing the testing benefits. It is showed by practices that the automatic test case generation method based on the evolving method can achieve the higher testing coverage rate while reducing the software testing cost.
The random resonance theory can be described as that: the input noise intensity is increased, and meanwhile the input signal intensity of the non-linear system is kept constant; when the noise intensity is at a specified level, the output signal-to-noise ratio of the system will increase at the output terminal of the non-linear system, and a single-peak (or multi-peak) resonance curve known by people in mechanics is generated. In recent years, with the adiabatic approximation, linear response port and intrinsic perturbation expansion as the main theories, the researches on the random resonance have made huge progress and gained wide applications in engineering field.
The software testing based on the evolving method generally excessively pursues the improvement of the software space coverage and emphasizes the characteristics, such as mutation and crossover, which highlight the diversity and jump out of the method local convergence. Thus, during the searching process, even if the information same or similar as the software fault phenomenon is temporarily searched, the test sequence uncorrelated to or far from the fault phenomenon is highly possible to be generated during the testing process of the next iterative evolution, which causes the too small sample size of the effective faults in the local space. Thus, when the local optimum or approximate solution occurs, the mutation operation which works normally should be paused, and a new strategy is nested, so as to fully utilize the local rare event which is beneficial and seldom appears (especially when the event is close to the fault phenomenon) for generating more test sequences close to the fault phenomenon and increasing the recurrence probability of the software fault. For the above situation, the inventors provide a construction method of a software accelerated testing technology based on a noise resonance. In the technology, the random resonance theory is used for reference that: the software to be tested is seen as a dynamic non-linear chaotic information system; various types of noises are exerted at the input spatial terminal of the program; the internal noise generated by the related coupling and interferences in the program space is utilized for coupling and transmission of the noise, so as to amplify the noise in the program space and the output space, trigger the noise resonance at the output terminal as far as possible, generate more test cases related to the fault, and realize the rapid recurrence of the software fault phenomenon or increase the recurrence probability. With the above technology, the noise is fully utilized, and the accelerated testing of the software is realized based on the noise resonance theory.
The software testing technology is constructed based on the noise resonance theory, and the software accelerated testing technology based on the noise resonance is formed, so that the credibility, safety and usability of the software are increased while the software testing efficiency is increased.